Reduce risk and ensure that the data has been migrated and transformed
Data Migrations have become one of the most challenging initiatives for IT managers. It is challenging whether businesses are migrating from:
- Legacy systems to a new system
- From one vendor’s software to another’s or
- From on-premises to the cloud
Data Migration refers to moving the Legacy (old) data to Target (new) data source. Data Warehousing being an important errand for all sort of businesses in regards of the cost and performance. A process known as ETL (Extract Transform Load) is initiated, that gathers data from CRMs, Data Servers or any flat files transformed it while staging area and collected in Data Warehouse. It might fail sometime and results in loss of information, data
Various types of Data Migration Testing are
Schema Compare Tests: Make sure that the data model or schema structure is matching between the source and target system. Users can easily query the metadata tables to pull the information for validating.
- Check if the table and column name is the same between source and target
- Datatype mapping between source and destination should be correct. Example source column with INT datatype should be NUMERIC in the target system
- Verify the views, primary keys, and indexes are also matching.
Row Count Tests: The most basic type of check is to make sure the count is for a table between source and target is matching.
- One time Row Count checks for the initial loads of all the tables
- Row Count checks for delta loads of all or specific tables
Data Comparison Tests: Compare the data in all of the tables, rows by row and column by column. This will certify that data migration was a success.
- Check the first name column in the source, and the target is the same.
- Ensure the date value is matching even though the format is different between the source and the target.
Data Aggregation Tests: Organizations can perform aggregated checks for really high volume tables between source and target. This is necessary as row by row comparison for billions of rows in a table can be costly.
Organizations can perform aggregated checks for really high volume tables between source and target. This is necessary as row by row comparison for billions of rows in a table can be costly.